A founder called us convinced his case was hopeless. Nothing about his company existed in English, he said: no coverage from an American trade outlet, no mention in an English-language wire report, nothing a reviewer skimming quickly would recognise. What he did have sat in a folder he almost didn't bring up: two feature profiles in his country's leading financial daily, a chapter in an industry history published by a national business school, and a long interview in a regional trade magazine. All of it in his own language. He had decided in advance that none of it would count.
It counts. The English-language Wikipedia is written in English, not exclusively about English-speaking subjects, and its sourcing guideline never asks what language a source is written in. It asks whether the source is independent, reliable, and gives the subject substantial coverage. A profile in a serious regional daily does the same work in Portuguese, Korean, or Finnish as it would in English.
English-language sources are preferred over non-English sources of equal quality and relevance, so that readers can easily verify the content. However, do not reject reliable non-English sources simply because they are not in English.
The preference for English is a courtesy to readers of an English-language project, not a filter on what may be true about a subject. When a source in another language is the strongest one available, it goes into the citation with a translation of the relevant passage attached, so that another editor, months or years later, can check that the source says what it is claimed to say without having to trust the person who added it.
That practice, translate and verify, is worth separating from a genuinely thin record. Sometimes what a client hands us is not a language problem but an independence problem: sponsored placements dressed as articles, a trade association's newsletter, a press release syndicated by a regional wire under a reporter's byline. Those fail on the same grounds in any language. Confusing the two, and assuming the fix is simply securing English mentions, sends a client chasing coverage that answers the wrong deficiency.
So we ask for everything, in whatever language it was written, before deciding anything is missing. Half the cases that arrive looking impossible turn out to be already built, just filed under the wrong problem.
